摘要： 提出了基于结构分类的手势识别方法,即在字母手势图像中寻找轮廓所包含的结构特征信息,并以此进行分类判断。这些结构特征中包含了大量的特征信息,如手指个数、手指间距离等,利用这些信息就可以运用传统的结构分类方法对手势加以识别。

关键词: 人机交互, 字母手势, 结构分类, 手势识别

Abstract: This paper proposes gesture recognition algorithm based on structure classification, which is to extract structure feature from the contour of alphabet gesture frames and then make the judge by classifying them. These structure feature includes massive information——number of fingers, distances between fingers, etc. By use of these information, the gesture can be recognized through traditional structure-classification method.

中图分类号: 

• TP391.4